Forum OpenACS Q&A: Re: error installing xotcl 1.4

Collapse
Posted by Richard Hamilton on
Bingo! Earlier error indeed located.

Yes the problem persists beyond server restart. It renders the server completely unable to return pages presumably because it is a request filter that is failing.

[15/Nov/2008:09:52:42][12341.65541][-default:0-] Error: Error sourcing /web/server/packages/xotcl-request-monitor/tcl/throttle_mod-procs.tcl:
invalid command name "thread::mutex"
while executing
"thread::mutex create"
(procedure "init" line 26)
::throttle ::xotcl::THREAD->init
::xotcl::THREAD ::xotcl::Class->create
invoked from within
"::xotcl::THREAD create throttle {
#
#set package_id [::xo::parameter get_package_id_from_package_key \
# -package_key "xotcl-..."
(file "/web/server/packages/xotcl-request-monitor/tcl/throttle_mod-procs.tcl" line 2)
invoked from within
"source $__file "

Collapse
Posted by Gustaf Neumann on
This message means that libthread is not correctly installed.

Check the installation steps in https://openacs.org/xowiki//tag/libthread

Have you compiled libthread with --with-aolserver? Have you added the ns_param?

Collapse
Posted by Richard Hamilton on
Gustaf,

Thank you very much indeed. I had not done so. In fact I had got up this morning specifically to re-trace my steps with the libthread library because having trawled through the code it seemed the most likely source of trouble. Your feedback saved me hours!!

I installed it and the following things happened:

1) There was an error associated with the ::thread object as before and an undeclared variable 'f'

2) After re-starting the server this error disappeared

3) The server was very much slower in returning pages than before the request monitor had been installed

4) When I selected xotcl-request-monitor from the applications list on the openacs index page, the server seemed to get stuck processing the request and no longer responded to any other requests either.

5) After uninstalling xotcl-request-monitor using the package manager the site returned to its previously brisk behaviour.

I have no idea why this is - perhaps my own omissions - but I thought that I would provide the feedback anyway.

Many Thanks
Richard

Collapse
Posted by Gustaf Neumann on
Not sure, what you mean by the "::thread object". Anyhow, i am not much concerned about error which disappear after restart.

Under normal conditions, one is not able to notice the difference in speed between using the request monitor or not (on my notebook the difference is less than 1ms). Something seems still quite wrong in your setup.